OBJECTIVES: To evaluate whether the use of exogenous melatonin affects sleep, reduces the prevalence of delirium, and decreases the need for analgosedation and to assess whether serum melatonin indices correlate with exogenous administration in critically ill patients. DESIGN: Double-blind, randomized, placebo-controlled study. SETTING: Multicenter ICUs of two tertiary hospitals. PATIENTS: A total of 203 adult patients who were admitted to the ICU and administered with analgesics and/or sedatives. INTERVENTIONS: Oral melatonin (10 mg) or placebo for up to seven consecutive nights. MEASUREMENTS AND MAIN RESULTS: The number of observed sleeping hours at night was assessed by the bedside nurse. Sleep quality was evaluated using the Richards Campbell Questionnaire Sleep (RCSQ). The prevalence of delirium, pain, anxiety, adverse reactions, duration of mechanical ventilation, length of ICU and hospital stays, and doses of sedative and analgesic drugs administered were recorded. The use of analgesics and sedatives was assessed daily. Melatonin levels were determined by enzyme-linked immunosorbent assay. Based on the RCSQ results, sleep quality was assessed to be better in the melatonin group than that in the placebo group with a mean (SD) of 69.7 (21.2) and 60.7 (26.3), respectively (p = 0.029). About 45.8% and 34.4% of participants in the melatonin and placebo groups had very good sleep (risk ratio, 1.33; 95% CI, 0.94-1.89), whereas 3.1% and 14.6% had very poor sleep (risk ratio, 0.21; 95% CI, 0.06-0.71), respectively. No significant difference was observed regarding the days free of analgesics or sedatives, the duration of night sleep, and the occurrence of delirium, pain, and anxiety. Melatonin serum peak levels at 2 AM were 150 pg/mL (range, 125-2,125 pg/mL) in the melatonin group and 32.5 pg/mL (range, 18.5-35 pg/mL) in the placebo group (p < 0.001). CONCLUSIONS: Melatonin was associated with better sleep quality, which suggests its possible role in the routine care of critically ill patients in the future.